ABSTRACT
ABSTRACT
The unicellular motile green alga Haematococcus lacustris (Chlorophyceae) is an economically important species capable of producing the high value pigment astaxanthin, used widely in aquaculture, pharmaceutical and cosmetic industries. Phylogenetic diversity in this genus is not very well understood even though significant physiological variability between strains has been observed. Understanding diversity can be important from a biotechnology perspective. Here we propose a new species of Chlorophyta, Haematococcus alpinus, isolated from Canyon Creek in the New Zealand alpine zone. This species is distinguished from others by location, number and thickness of cytoplasmic strands. Phylogenetic analysis based on 18S rDNA, ITS rDNA, rbcL genes and comparison of secondary structures of ITS1 showed that our species is distinct with no known close relatives.
Introduction
The ‘blood rain alga’, most commonly known in the scientific literature as Haematococcus pluvialis Flotow (Goksan et al. 2010), is one of the most important species in microalgal biotechnology because it can accumulate large amounts of the commercially valuable astaxanthin (3,3′-dihydroxy-β, β-carotene-4-4′-dione) pigment when stressed (Boussiba 2000). Astaxanthin has important applications in nutraceutical and cosmetics industries as a potent antioxidant (Kobayashi & Sakamoto 1999; Guerin et al. 2003). Haematococcus is not the only natural source of this pigment, but it is by far the most potent producer, accumulating up to 2.5%–4% of its dry weight as astaxanthin (Li et al. 2011; Han et al. 2012).
The type species of Haematococcus has been the subject of a long-standing nomenclatural issue, in which it has been confusingly referred to by two names: H. lacustris (Girod-Chantrans) and H. pluvialis Flot. This issue with synonymy has recently been resolved with the current name for the widely cited H. pluvialis being established as H. lacustris (Girod-Chantrans) Rostafinski (Nakada & Ota 2016). We follow this taxonomy hereafter.
Several new strains with interesting properties, but still identified as H. lacustris, were isolated in recent years. Chekanov et al. (2014) reported a novel Arctic strain with elevated salt tolerance (up to 25%). Klochkova et al. (2013) discovered a cold tolerant strain that was capable of growing and producing astaxanthin at low temperatures (4–10 °C). Other prior studies have also documented significant physiological variability between strains, with up to threefold differences in biomass production, growth rate and/or astaxanthin content when grown under identical conditions (González et al. 2009; Zhang et al. 2009; Mostafa et al. 2011; Su et al. 2014). These findings suggest that there may be undiscovered species diversity within strains known as H. lacustris (Allewaert et al. 2015). Understanding this diversity is important from a biotechnology perspective as it provides opportunities for selection of fast growing, productive species and strains that are adapted to local growing conditions and for a reduction of economic and environmental costs in industries.
The genus Haematococcus is principally characterised by the unique relationship between the cell wall and the protoplast. These two structural elements are connected by strands of cytoplasm that differ in thickness and degree of branching, depending on the species (Smith 1933; Fritsch 1935; Droop 1956). Using these characters, Droop (1959) established the genus Balticola and transferred H. droebakensis and H. buetschlii to it, a decision rejected by Pocock (1960) and Almgren (1966) but subsequently confirmed using phylogenetic reconstruction based on DNA sequences (Buchheim et al. 2013). Other recent studies have contributed further knowledge of phylodiversity in this group of algae, with Pegg et al. (2015) demonstrating a close relationship between the coccoid species Ettlia carotenosa and Haematococcus, and Allewaert et al. (2015) establishing two new species of Haematococcus from Europe.
Here we propose a novel species of Haematococcus isolated from the alpine zone in Canyon Creek, New Zealand, which differs from previously documented species according to morphological and phylogenetic criteria.
Materials and methods
Collection site
A sample of green-coloured water and benthos (approximately 1–2 g) was collected from a small pool on a rock outcrop, directly into a sterile 9 mL polycarbonate test tube, at an altitude of 1700 m in upper Canyon Creek, Ahuriri Conservation Park, South Island, New Zealand (44°10'58.91ʺS, 169°34'39.59ʺE) on 16 February 2011. After collection the specimen was refrigerated overnight at a field camp and for 2 days in the laboratory before inoculation of cultures.
Culturing
The unialgal strain was isolated as described previously (Gopalakrishnan et al. 2014) using 5.5 cm diameter Petri plates containing 1.2% agarised BG-11 medium (Rippka et al. 1979) diluted to 10% concentration by aseptic spread plate technique in a laminar flow hood. The agar was first prepared at double strength and washed for 5 days after autoclave sterilisation by soaking daily in fresh sterile distilled water. Media and agar were autoclaved separately, cooled to c. 45 °C, and recombined before pouring onto plates. After inoculation the Petri plates were incubated at 5–10 °C with light intensity of approximately 50 µmol photons m−2 s−1 for 3 weeks. After 3 weeks, each distinct colony type on each plate was examined microscopically and isolated into unialgal culture by aseptic streak plate technique.
Light microscopic analysis
Specimens were examined under a light microscope (Leica DMLB) and images collected using a Canon DS126271 digital camera system.
DNA isolation and sequencing
Plates containing unialgal cultures were checked for contamination microscopically. DNA was extracted from cultures using paramagnetic particle technology, Automated Maxwell 16 Instruments (Promega Corporation), and the resulting samples processed using Zymo Clean and Concentrate columns, according to the manufacturer’s instructions (Zymo Research). The extracted DNA was amplified in PCR using primers for the 18S rDNA gene (Hoham et al. 2002) and the rbcL gene (Nozaki et al. 1995). The ITS rDNA region was amplified by using the universal primers ITS4 (5′- TCCTCCGCTTATTGATATGC-3’)(White et al. 1990) and 1800F (5′-ACCTGCGGAAGGATCATTG-3’) (Friedl 1996). All amplified products were visualised using agarose gel electrophoresis and ethidium bromide staining. The amplified products were diluted prior to using in Big Dye Terminator 3.1 sequencing reactions, and capillary separation of the products was carried out by Landcare Research, Auckland, New Zealand. Electropherograms were checked using Sequencher 4.8 (Gene Codes Corporation).
Phylogenetic analysis
Three separate sequence datasets containing 18S rDNA, ITS rDNA (including ITS1, 5.8S and ITS2) and rbcL were aligned using MEGA v6 (Tamura et al. 2011) and checked by eye. The outgroups in the rbcL and ITS datasets were chosen based on Allewaert et al. (2015). The ITS dataset consisted of 95 sequences containing 750 sites, including 189 variable sites and 137 parsimony informative sites. The rbcL dataset consisted of 23 sequences containing 1386 sites, including 288 variable and 189 parsimony-informative sites. The 18S rDNA dataset consisted of 27 sequences containing 1171 sites, of which 99 sites were variable and 17 parsimony informative.
Analyses were carried out using two methods. The first was the program MrBayes v3.1.2B4 (Ronquist & Huelsenbeck 2003) constructing a Bayesian analysis of phylogeny using MEGA to select a model of DNA substitution. The K2 model was used for 18S rDNA, K2 + G for ITS and GTR + G for rbcL genes.
Two independent runs of 1 million generations were used; each with 4 chains and random starting trees, and a consensus tree was constructed discarding the first 500,000 generations as burn-in (judged from log-likelihood plots). Means of parameter estimates were compared with their associated variances to assess effective modelling and the efficiency of chain swapping was evaluated using the program output.
The second method was a maximum-parsimony full heuristic bootstrap (MPB) analysis implemented in PAUP 4.0b10 (Swofford 2002) employing the following settings: branches collapsed if maximum length equals 0; DELTRAN character state optimisation; and assignment of character states not observed in terminal taxa allowed at internal nodes. Nonparametric bootstrap values for nodes were calculated on the basis of 1000 resamplings.
In addition, a comparison of the ITS1 sequence of our strain with that modelled for the epitype of H. lacustris by (Nakada & Ota 2016) () was carried out, using mfold (Zuker et al. 1999). ITS2 was omitted due to low strain coverage.
Table 1. Secondary structure of helices I and II from a representative genetic diversity of Haematococcus strains, based on the lectotype strain illustrated by Nakada & Ota (2016). Red letters indicate non-compensatory substitutions, blue represent insertions and green represent compensatory base changes (CBCs) (all relative to the sequence of the type).
Taxonomy
Haematococcus Flot., Beobachtungen über Haematococcus pluvialis. Verhandlungen der Kaiserlichen Leopoldinisch-Carolinischen Deutschen Akademie der Naturforscher 12(Abt. 2), p. 413 (1844)
Type strain. Haematococcus lacustris (Girod-Chantrans 1802) Rostaf. 1875
Haematococcus alpinus Mazumdar et Gopalakrishnan sp. nov. Figure 1
Holotype. New Zealand, Canterbury, Upper Canyon Creek, K. Gopalakrishnan, 16 February 2011, CHR 554372 (preserved specimen from cultured material).
Habitat. Green coloured rain pool, 1700 m altitude.
New Zealand localities: upper Canyon Creek (this study).
Etymology. ‘alpinus’, referring to the alpine region from where it was isolated.
Published online:
27 April 2018Figure 1. Haematococcus alpinus (photographed from culture LCR-26-CC-1f). A, Vegetative cell in motile stage; B, non-motile vegetative cell; C, asexual reproduction in a vegetative cell; D, cyst cell with pigment accumulation. A–D: light micrographs; E transmission electron micrograph. Scale bars A–D = 10 µm; E = 5 µm. c, cytoplasmic strands; e, envelope; f, flagella; p, papilla; pi, pigment; py, pyrenoid; t, thylakoid membranes.
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/tnzb20/2018/tnzb20.v056.i02/0028825x.2018.1458737/20180620/images/medium/tnzb_a_1458737_f0001_c.jpg"})
Figure 1. Haematococcus alpinus (photographed from culture LCR-26-CC-1f). A, Vegetative cell in motile stage; B, non-motile vegetative cell; C, asexual reproduction in a vegetative cell; D, cyst cell with pigment accumulation. A–D: light micrographs; E transmission electron micrograph. Scale bars A–D = 10 µm; E = 5 µm. c, cytoplasmic strands; e, envelope; f, flagella; p, papilla; pi, pigment; py, pyrenoid; t, thylakoid membranes.
Description
Vegetative cells spherical to ovoid in shape, 10–20 µm in diameter, 20–30 µm in length, biflagellate, with thin cell walls separated from the protoplast by a wide space filled with a slimy substance (Figure 1A). Cell membrane is narrow. Protoplast is mostly rounded at the apex but occasionally pyriform. Cytoplasmic strands are usually branched, distributed uniformly around the protoplast and connecting the protoplast to the cell wall, but sometimes limited to the posterior side (opposite the flagella). Chloroplast single, cup-shaped, parietal, occupying 1/3 of the protoplast and containing 2–6 scattered spherical pyrenoids (Figure 1B, E). Two isokont anterior flagella are present, as long as the cell and surrounded by short divergent tubes between the protoplast and outer envelope. Nucleus centrally located. Cysts characterised by flagella loss, gradual increase in size and development of thick cell wall. Resting cells contain high concentrations of blood-red pigment (Figure 1D). Reproduction by formation of zoospores containing 2–4 cells per sporangium (Figure 1C). Sexual reproduction occasionally observed in some cells under stressed conditions, isogamous, resulting in quadriflagellate motile zygotes that subsequently lose their flagella and develop thickened walls.
Molecular data
Sequences of the rbcL (Figure 2), ITS (Figure 3) and 18S rDNA (Figure S1) regions placed the species on isolated long branches with no close relatives. Furthermore, the secondary structure of two helices of the ITS1 region showed marked differences () with other species. The rbcL sequences suggest H. rubicundus Allewaert & Vanormelingen (eight strains) as the closest relative to our strain, although this was only robustly supported by the MP analysis. The lowest P distance separating our strain from H. rubicundus strains was 0.033. However, according to the ITS rDNA sequences our strain formed a sister relationship with five strains of H. lacustris (HP075, HP099, HP060, HP036, HP038); two of which (HP036 and HP099) also formed ‘lineage B’ in the phylogeny presented by Pegg et al. (2015).
Published online:
27 April 2018Figure 2. Phylogenetic analysis of selected Haematococcus strains inferred from rbcL sequences. The tree topology is that inferred using MrBayes v3.1.2B4. Values above branches correspond to Bayesian posterior probabilities/Maximum Parsimony bootstrap values. Scale bar represents 0.1 changes/site. The type strain of Haematococcus lacustris is UTEX 16 (Nakada & Ota 2016).
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/tnzb20/2018/tnzb20.v056.i02/0028825x.2018.1458737/20180620/images/medium/tnzb_a_1458737_f0002_b.gif"})
Figure 2. Phylogenetic analysis of selected Haematococcus strains inferred from rbcL sequences. The tree topology is that inferred using MrBayes v3.1.2B4. Values above branches correspond to Bayesian posterior probabilities/Maximum Parsimony bootstrap values. Scale bar represents 0.1 changes/site. The type strain of Haematococcus lacustris is UTEX 16 (Nakada & Ota 2016).
Published online:
27 April 2018Figure 3. Phylogenetic analysis of selected Haematococcus strains inferred from ITS rDNA sequences. The tree topology is that inferred using MrBayes v3.1.2B4. Values above branches correspond to Bayesian posterior probabilities/Maximum Parsimony bootstrap values. Scale bar represents 0.05 changes/site. The type strain of Haematoccocus lacustris is encompassed within the H. lacustris clade indicated.
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/tnzb20/2018/tnzb20.v056.i02/0028825x.2018.1458737/20180620/images/medium/tnzb_a_1458737_f0003_b.gif"})
Figure 3. Phylogenetic analysis of selected Haematococcus strains inferred from ITS rDNA sequences. The tree topology is that inferred using MrBayes v3.1.2B4. Values above branches correspond to Bayesian posterior probabilities/Maximum Parsimony bootstrap values. Scale bar represents 0.05 changes/site. The type strain of Haematoccocus lacustris is encompassed within the H. lacustris clade indicated.
Results of secondary structure modelling using mfold (Zuker et al. 1999) suggested that the nearest relative of H. alpinus among those analysed is the type strain of H. laustris, by virtue of the fact that the two are separated by no CBCs, unlike the other species ().
Discussion
In this study we have reported a new species of Haematococcus from New Zealand alpine regions, including comparison of its sequence data with those of other Haematococcus strains. We sequenced the complete ITS rDNA region as it provides more phylogenetic information than 1TS2 alone (Allewaert et al. 2015). Morphological differences to published descriptions of species, and ITS rDNA and rbcL genes, indicate that our strain LCR-26C-1f is a species of Haematococcus new to science.
The nomenclatural situation of H. pluvialis and H. lacustris (Girod-Chantrans) Rostafinski, the two vying names for the type species of Haematococcus, has proven to be controversial, largely due to inadequate type descriptions. As pointed out by Nakada & Ota (2016), many phycologists who accept the synonymy of these names have prioritised H. pluvialis (Ettl & Komárek 1982; Melkonian & Preisig 2000; Massjuk et al. 2011; Buchheim et al. 2013). Nakada & Ota (2016) recently formalised this synonymy and designated the correct name for the type of Haematococcus as H. lacustris, not H. pluvialis.
Prior to our investigation, the genus Haematococcus contained four valid species: H. lacustris; H. rubicundus; H. rubens Allewaert & Vanormelingen; and H. carocellus R.H. Thompson & D.E. Wujek. Other published species of Haematococcus (H. buetschlii Blochmann, H. capensis Pocock, H. droebakensis Wollenweber. and H. zimbabwiensis Pocock) were transferred to Balticola Droop by Buchheim et al. (2013). Although cells of both genera exhibit cytoplasmic strands connecting the cytoplasm to the outer envelope, those in Balticola are ‘robust’ (i.e. thickened at the base) (Buchheim et al. 2013), whereas those of Haematococcus are ‘delicate’, of a constant narrow width. A variant of the robust morphology is also found in H. carocellus (Thompson & Wujek 1989), which so far lacks molecular data, leading Buchheim et al. (2013) to suggest that this species belongs in Balticola also. The cytoplasmic strands found in H. alpinus conform to the delicate morphotype (Figure 1A). Although the species Haematococcus capensis has been transferred to Balticola, to our knowledge its varieties have not been. One of these is from New Zealand: H. capensis var. novae-zelandiae Pocock (Broady et al. 2012). Unfortunately we were unable to find a description or illustration of this variety, and it also lacks molecular data. However, since it was described by the same author (Pocock) as the type variety of the species, and that species is now placed in Balticola (Buchheim et al. 2013), the placement of var. novae-zelandiae in Balticola would also be highly likely.
According to Buchheim et al. (2013): ‘Of all the taxa historically ascribed to the genus Haematococcus, only H. pluvialis [now H. lacustris] produces a vegetative resting stage (the akinete) that accumulates copious amounts of the carotenoid astaxanthin’ (318). This feature is also lacked by Balticola, species of which are variously yellow-red (Blochmann 1886), reddish ochre (Wollenweber 1907) and deep golden-brown (Pocock 1960). Cysts of B. zimbabwiensis (Pocock) Droop and H. carocellus so far have not been reported (Pocock 1960; Thompson & Wujek 1989). Haematococcus alpinus can additionally be distinguished from H. lacustris based on pyrenoid number, protoplast shape and cell length. One or two (sometimes several) pyrenoids, an ovoid or ellipsoidal protoplast and a cell length of 29–39 µm is found in the latter species (Girod-Chantrans 1802).
The remaining two species of Haematococcus can also be distinguished from H. alpinus using morphology. In H. rubicundus, cytoplasmic strands are usually absent and the number of pyrenoids ranges from 1–11 (Allewaert et al. 2015). In H. alpinus the cytoplasmic strands are almost always present and the number of pyrenoids ranges from 2–6. The protoplast of H. rubens is predominantly pyriform and cytoplasmic strands when present are thicker than those found in H. alpinus.
In addition to morphological differences with other described species, H. alpinus clearly differs at the molecular level (Figures 2–3). The nearest relative of H. alpinus among these species is unclear; rbcL data suggest H. rubicundus (Figure 2), but other less well characterised strains (tentatively labelled H. lacustris lineage B) for which ITS data are available, appear to be closer relatives. Secondary structure models () tentatively support a conclusion that the type strain of H. lacustris is the closest relative of H. alpinus among the strains analysed, since H. alpinus is the only one to lack CBCs separating it from the type. However, the type strain is not part of lineage B.
 |
Supplementary Figure 1
Download JPEG Image (1932 KB)
Supplementary Figure and Tables
Download MS Word (391 KB)Related Research Data
Acknowledgements
We thank Quang A. Dang for his help with Adobe Photoshop and Dr Peter A. Gostomski (University of Canterbury) for his support and guidance.
Disclosure statement
No potential conflict of interest was reported by the authors.